Explaining the plot:
We see the self similarity matrix of the top four songs in the Dutch top 50 playlist on Spotify. Via Timbre we can compare the structure of the songs internally and find the sections that stand out the most. Also we can compare the four songs among each other.
The black squares (structure)
A typical structure for a pop song is as follows: intro, verse — chorus — verse — chorus — bridge — chorus, outro
If we look at the timbre based self-similarity matrices displayed in the plot we can make out the structure of the song along the black squares that form where the x and y axes meet. A larger black square means that part of the song (intro, chorus, verse, etc.) was consistent for a larger portion. This explains why the black squares are proportionate to each other as they are musically expressed as a multitude of “bars”. For instance, in the Number 1 hit “Amsterdam”, the into is 1 bar followed by a 4 bar chorus. As seen in the plot the black square representing the intro is 4 times smaller than the next.
The brightest squares (the outliers)
Now that we have a way of interpreting the structure, lets take a look at the brightest squares. These indicate the magnitude with which the timbre differs from that particular part of the song to other parts of the song. If you take a glance at all four of the plots you might notice that all plots read the highest difference in timbre between the intro and the bridge. This can be explained by listening to the songs, the intro of all four songs is a toned down version of the verse or chorus. With toned down I mean elements are missing: the beat, the vocals or the chords. These elements are introduced after the intro creating a change in timbre. Now lets look at the bridge, recognizable in the plot by its presence as a cross in the top right of each matrix. The bridge functions as a structural change in the song, and is defined by the means to provide contrast. Therefore it isn’t strange that the bridge is often the most outstanding structure. Notably, the highest contrast in three of the songs is found between the bridge and the intro. The only exeption is: “Je Blik Richting Mij”, where the intro is obviously the most outstanding structure and the bridge seems to bear the closest resemblance to the intro. Another notable outlier is the song “Amsterdam” where there seems to be two bridges. The two outliers are plotted in a chroma self-similarity matrix on the next page so we can take a closer look at the pitch information to see if we can find a reason for their divergence.
Explaining the plot:
If we look at the chromagram of both songs, it immediately becomes clear that there is a lot of repetition.
Explaining the plot:
The plot shows 4 top 50 playlist from European countries on the x axis: danceability and on the y axis: Speechiness. Color represents the minor and major scale use per track. I was hoping to find a more obvious relation between the x and y parameters but there are already some interesting findings. For instance, the use of minor scale in Dutch music indicates an increase of speech used in the songs, a similar but slightly weaker correlation is found in the plot of Italy’s top 50. Germany scores higher overall with the amount of speech used in songs whilst Spain has a remarkably dense cluster of danceable music.
Explaining the plot:
We isolate the number one songs of the Dutch and Spanish top 50 playlists: “Amsterdam” and “Cáyo La Noche”. Neither song reach a vocal part in the first ten seconds, due to the pitch varying nature of human vocals this leaves a cleaner chromagram to look at. When listening to “Amsterdam” we hear a synthesizer chord struck in an alternating pattern. The g chord is alternated with D# for the first bar and with a D chord in the second bar, this pattern is repeated throughout the song and gives off a strong minor key vibe due to the semitone distance between D and D#. The pattern is broken in the third bar for the singer to introduce his name, the instruments drop and only vocals are heard for 2 seconds. This is hard for the chromagram to visualize, therefore a stripe of activation is observed throughout the whole spectrum. Immediately after the vocals a dubble woodblock click reintroduces the beat, we can see the woodblock is C#.
The song “Cáyo La Noche” starts with piano chords C D# D. We can see the chords drawn out in the chromagram with the root nodes as brightest. Due to a vocalist the second half of the chromagram is murky. Also note that the two songs are coincidentally in the same key (look at the semitone difference between G G# and D D#). This raises the question is the C minor scale popular in all top 50 tracks?
Explaining the plot:
The number one song in the Spanish top 50 is one of many remixes. In the plot we see the original “Cayo De La Noche” ploted against one of the remixes called “Cayo De La Nocheh”. I’ve isolated the first 20 seconds for clarity. We can see the first piano chords clearly in the bottom left and can immediately tell the remix is a faster pace than the original. This leads the diagonal line (where the two songs meet) to be steep. Also the remix holds less speech than the original leaving the darker gaps along the Y axes.